Last Thursday an item crossed my news feed entitled The CDC's Mask Mandate Study: Debunked. We here at Crona-zona are generally pro-mask; however, we are also open to persuasion. So, I decided to look into the claims made in the article. Its target is the CDC's Morbidity and Mortality Monthly Report published in February, which concluded that mask mandates reduced COVID-19 hospitalizations.
The third paragraph offers sarcastic faux praise of the CDC for publishing an erratum correcting the initial version of the report. It was admirable that the CDC did this, said the author, because the first version was "plagued with inaccuracies."
Even with corrections, the report is no good, according to the fourth paragraph: "En face, CDC’s conclusion on mandates might appear to make sense unless one is familiar with the scientific data pertaining to the ineffectiveness of masking for prevention of the spread of Covid-19." In other words, it makes sense, but only if you're ignorant.
The author then provides a numbered set of—count 'em—15 links to scientific data proving that masks don't prevent virus spread. Wow! That sure is a lot of evidence! What is actually there?
- Six links were to opinion pieces, not to scientific research. Four of these were published by the same outlet, the American Institute for Economic Research (more on them later). Of these, one was by the same author as the piece under review, making the same arguments. One was a duplicate link.
- Four were links to reviews, meta-analyses (studies of studies), or reports that summarize but do not present new scientific data.
- Of the nine non-opinion pieces, six were published July 2020 or earlier, when less was known about the pandemic.
- One doesn't mention masks at all except in a reference. Another only has one paragraph on masks in a larger summary of research.
- Only four present actual new scientific data, and one of them is unpublished.
So, what about those scientific data papers? One, in
The Lancet, is a cohort study of COVID-19 transmission in Spain. It doesn't use the work "mask" except in one reference. The study is not about masks.
Another study looks at transmission among marine recruits during quarantine
. It states that all participants were required to wear masks. It can't draw any conclusions about the effect of mask wearing because there is no control condition. Accordingly, it doesn't draw any mask-related conclusions,
at all.
The author wants to conclude that because 0.9 % of recruits were infected when coming into a congregate setting and 1.3% were infected after a week that included mask wearing, that mask wearing doesn't work. They fail to mention that the recruits came in from two weeks of home quarantine and entered a setting where they double bunked, shared bathrooms, ate in dining halls, etc. That the increase was only 0.4% is remarkable.
Another study, the unpublished one, is a quasi-experiment. That means it doesn't manipulate control and treatment groups. Instead it tries to find existing differences in the population that can be compared. In this case, they looked at epidemiological data in different areas of Europe and compared that with varying mitigation strategies (including mask mandates) in the same areas. Importantly, they didn't actually measure mask-wearing.
The authors noted that "there was substantial heterogeneity in how the wearing of face coverings in the community was encouraged or mandated and in what contexts." They concluded with a caveat that, "[o]ur results on face coverings should be considered to be preliminary because the use of coverings was recommended or required only relatively late in the epidemics in each European country."
The
third study is a randomized controlled trial of Danish mask wearers. This one is a little more on-target. It randomly assigned participants to a masks-recommended group, and others to a control, and compared infection rates between the two. The bottom line is that there was a difference in favor of the mask group, but it was only a 0.3 percent reduced chance of being infected.
Yet this study isn't the smoking gun our debunker would have us believe. A
critique of the so-called DANMASK-19 study appeared in the same journal. It said the study's setting limits its statistical power and generalizability,
it used inaccurate antibody tests to measure infections, and only 46% of the treatment group reported adhering the mask-wearing guidelines.
Another critique said that while you expect the study to conclude that masks are useless, it shows the opposite (to the extent that it shows anything): The trial is inconclusive rather than negative, and it points to a likely benefit of mask wearing to the wearer—it did not examine the wider potential benefit of reduced spread of infection to others—and this even in a population where mask wearing isn’t mandatory and prevalence of infection is low.
The debunker goes on at length criticizing choices made by the CDC in designing their study, but continues citing these same 14 references.
What about the outlet where the debunking article appears? The American Institute for Economic Research is rated by Media Bias/Fact Check as a center-right, libertarian outlet with a "mixed record" of factual reporting. It says that another story (not by this same author), which claims that the asymptomatic spread of COVID-19 never occurs, is factually inaccurate, cherry-picks facts, and misinterprets findings of research.
To summarize, of fourteen references (remember, one was a duplicate) made in this article, all but four are opinion or review pieces. Of the four actual studies, one doesn't use the work "mask" except in a reference, one doesn't manipulate mask wearing, one is observational research that studies mask policies, and one has received convincing, published peer criticism.
The author implies that all 15 links present "scientific data," which they do not. The scientific data that is presented in four of the links does not support the author's claim, and the author fails to consider
evidence against his position. Furthermore, there is reason to question the credibility of the AIER outlet as a source on pandemic-related topics, and the author seems to be pushing an anti-mask agenda.
So do Masks Work?
An important point is that it is basically impossible to produce the conclusive evidence of mask effectiveness that our debunker seems to want. You just can't do an adequate randomly controlled trial of the effectiveness of masks in preventing transmission in naturalistic settings.
There are too many other factors that could impact infection chances (other mitigation measures, community infection rates, individual differences in risky behavior, undetected prior infections conferring immunity, etc.) to be controlled. Even if you could control these things, you couldn't surveil participants in the treatment group to ensure that they wear masks, or in the control group to ensure that they don't.
If we accept the idea that COVID-19 is caused by transmission of respiratory droplets, there is plenty of evidence that masks stop this. You can
see it for yourself. There are
studies proving that masks arrest aerosol particles the same size as those spreading SARS-CoV-2. Where masks fail it's largely because of leakage, something that
can be fixed. So, by applying common sense we can conclude that, yes, masks work. If masks work, then mask mandates are a good idea.
It might be that you're a libertarian who, because of
reactance, lashes out at any regulation of personal behavior. But for the rest of us, and as a matter of public health policy, there is no evidence that masks cause harm so long as you don't have breathing problems. Meanwhile there is plenty of evidence that they work. Nothing in the article reviewed here "debunks" that conclusion.
